In [1]:
from skimage.feature import hog
from sklearn.svm import LinearSVC
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from scipy.ndimage.measurements import label
import AdvancedLaneLines as lanes
import matplotlib.image as mpimg
import matplotlib.pyplot as plt
from moviepy.editor import VideoFileClip
from IPython.display import HTML
import numpy as np
import pickle
import cv2
import glob
import time
from tqdm import tqdm
%matplotlib inline
Utility functions¶
Mostly used from lesson as is, but there are some changes and modifications like changes in default parameters due to deprecation warnings, adjyustment some parameters etc.
In [2]:
def get_hog_features(img, orient, pix_per_cell, cell_per_block,
vis=False, feature_vec=True):
# Call with two outputs if vis==True
if vis == True:
features, hog_image = hog(img, orientations=orient,
pixels_per_cell=(pix_per_cell, pix_per_cell),
cells_per_block=(cell_per_block, cell_per_block),
transform_sqrt=False,
visualise=vis, feature_vector=feature_vec)
return features, hog_image
# Otherwise call with one output
else:
features = hog(img, orientations=orient,
pixels_per_cell=(pix_per_cell, pix_per_cell),
cells_per_block=(cell_per_block, cell_per_block),
transform_sqrt=False,
visualise=vis, feature_vector=feature_vec)
return features
# Define a function to compute binned color features
def bin_spatial(img, size=(32, 32)):
# Use cv2.resize().ravel() to create the feature vector
features = cv2.resize(img, size).ravel()
# Return the feature vector
return features
# Define a function to compute color histogram features
# NEED TO CHANGE bins_range if reading .png files with mpimg!
def color_hist(img, nbins=32, bins_range=(0, 256)):
# Compute the histogram of the color channels separately
channel1_hist = np.histogram(img[:,:,0], bins=nbins, range=bins_range)
channel2_hist = np.histogram(img[:,:,1], bins=nbins, range=bins_range)
channel3_hist = np.histogram(img[:,:,2], bins=nbins, range=bins_range)
# Concatenate the histograms into a single feature vector
hist_features = np.concatenate((channel1_hist[0], channel2_hist[0], channel3_hist[0]))
# Return the individual histograms, bin_centers and feature vector
return hist_features
Extract features from training images (Car and Non-Car)¶
Modified by using only HOG features as most efficient (test images have different colors, brightness, color deepness)
In [3]:
# Define a function to extract features from a list of images
# Have this function call bin_spatial() and color_hist()
# Leave only HOG parameters
def extract_features(imgs, color_space='RGB', orient=9, pix_per_cell=8, cell_per_block=2, hog_channel=0):
# Create a list to append feature vectors to
features = []
# Iterate through the list of images
images_pbar = tqdm(imgs, 'Extracting features', unit='images')
for file in images_pbar:
# Read in each one by one
image = mpimg.imread(file)
# apply color conversion if other than 'RGB'
if color_space != 'RGB':
if color_space == 'HSV':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2HSV)
elif color_space == 'LUV':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2LUV)
elif color_space == 'HLS':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2HLS)
elif color_space == 'YUV':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2YUV)
elif color_space == 'YCrCb':
feature_image = cv2.cvtColor(image, cv2.COLOR_RGB2YCrCb)
else: feature_image = np.copy(image)
# Call get_hog_features() with vis=False, feature_vec=True
if hog_channel == 'ALL':
hog_features = []
for channel in range(feature_image.shape[2]):
hog_features.append(get_hog_features(feature_image[:,:,channel],
orient, pix_per_cell, cell_per_block,
vis=False, feature_vec=True))
hog_features = np.ravel(hog_features)
else:
hog_features = get_hog_features(feature_image[:,:,hog_channel], orient,
pix_per_cell, cell_per_block, vis=False, feature_vec=True)
# Append the new feature vector to the features list
features.append(hog_features)
# Return list of feature vectors
return features
Loading training data¶
In [4]:
#Read in cars and notcars
car_images = glob.glob('classify-images/vehicles/**/*.png')
notcar_images = glob.glob('classify-images/non-vehicles/**/*.png')
print('Found car images: ' + str(len(car_images)))
print('Found non car images: ' + str(len(notcar_images)))
Some training data visualization¶
In [5]:
fig, ax = plt.subplots(6, 6, figsize=(12, 12))
ax = ax.ravel()
for i in range(36):
img = mpimg.imread(car_images[np.random.randint(0, len(car_images))])
ax[i].axis('off')
ax[i].set_title('Car', fontsize=10)
ax[i].imshow(img)
In [6]:
fig, ax = plt.subplots(6, 6, figsize=(12, 12))
ax = ax.ravel()
for i in range(36):
img = mpimg.imread(notcar_images[np.random.randint(0, len(notcar_images))])
ax[i].axis('off')
ax[i].set_title('Not-Car', fontsize=10)
ax[i].imshow(img)
HOG extract features visualization¶
In [7]:
#Select random image
fname = car_images[np.random.randint(0, len(car_images))]
name = fname.split('\\')[-1].split('.')
img = mpimg.imread(fname)
hog_img = cv2.cvtColor(img, cv2.COLOR_RGB2YCrCb)
_, dst1 = get_hog_features(hog_img[:,:,0], 9, 8, 2, vis=True, feature_vec=True)
_, dst2 = get_hog_features(hog_img[:,:,1], 9, 8, 2, vis=True, feature_vec=True)
_, dst3 = get_hog_features(hog_img[:,:,2], 9, 8, 2, vis=True, feature_vec=True)
# Visualize
f, (ax1, ax2, ax3, ax4) = plt.subplots(1, 4, figsize=(16, 5))
f.subplots_adjust(hspace = 1, wspace=1.2)
ax1.imshow(img)
#mpimg.imsave('output_images/car_original.' + name[1], img, cmap='gray')
ax1.set_title('Source Image', fontsize=14)
ax2.imshow(dst1, cmap='gray')
#mpimg.imsave('output_images/car_hog1.' + name[1], dst1, cmap='gray')
ax2.set_title('HOG[0] Features', fontsize=14)
ax3.imshow(dst2, cmap='gray')
#mpimg.imsave('output_images/car_hog2.' + name[1], dst2, cmap='gray')
ax3.set_title('HOG[1] Features', fontsize=14)
ax4.imshow(dst3, cmap='gray')
#mpimg.imsave('output_images/car_hog3.' + name[1], dst3, cmap='gray')
ax4.set_title('HOG[2] Features', fontsize=14)
Out[7]:
In [8]:
#Select random image
fname = notcar_images[np.random.randint(0, len(notcar_images))]
name = fname.split('\\')[-1].split('.')
img = mpimg.imread(fname)
hog_img = cv2.cvtColor(img, cv2.COLOR_RGB2YCrCb)
_, dst1 = get_hog_features(hog_img[:,:,0], 9, 8, 2, vis=True, feature_vec=True)
_, dst2 = get_hog_features(hog_img[:,:,1], 9, 8, 2, vis=True, feature_vec=True)
_, dst3 = get_hog_features(hog_img[:,:,2], 9, 8, 2, vis=True, feature_vec=True)
# Visualize
f, (ax1, ax2, ax3, ax4) = plt.subplots(1, 4, figsize=(16, 5))
f.subplots_adjust(hspace = 1, wspace=1.2)
ax1.imshow(img)
#mpimg.imsave('output_images/not_car_original.' + name[1], img, cmap='gray')
ax1.set_title('Source Image', fontsize=14)
ax2.imshow(dst1, cmap='gray')
#mpimg.imsave('output_images/not_car_hog1.' + name[1], dst1, cmap='gray')
ax2.set_title('HOG[0] Features', fontsize=14)
ax3.imshow(dst2, cmap='gray')
#mpimg.imsave('output_images/not_car_hog2.' + name[1], dst2, cmap='gray')
ax3.set_title('HOG[1] Features', fontsize=14)
ax4.imshow(dst3, cmap='gray')
#mpimg.imsave('output_images/not_car_hog3.' + name[1], dst3, cmap='gray')
ax4.set_title('HOG[2] Features', fontsize=14)
Out[8]:
Extract features¶
Extracting HOG features from train image sets, split sets into train and test sub-sets
In [9]:
colorspace = 'YUV' # Can be RGB, HSV, LUV, HLS, YUV, YCrCb
orient = 11
pix_per_cell = 16
cell_per_block = 2
hog_channel = 'ALL' # Can be 0, 1, 2, or "ALL"
t = time.time()
car_features = extract_features(car_images, color_space=colorspace, orient=orient,
pix_per_cell=pix_per_cell, cell_per_block=cell_per_block,
hog_channel=hog_channel)
notcar_features = extract_features(notcar_images, color_space=colorspace, orient=orient,
pix_per_cell=pix_per_cell, cell_per_block=cell_per_block,
hog_channel=hog_channel)
t2 = time.time()
print(round(t2-t, 2), 'Seconds to extract HOG features...')
X = np.vstack((car_features, notcar_features)).astype(np.float64)
# Fit a per-column scaler - only if used several types of features like HOG, color histogramm, spatial
#X_scaler = StandardScaler().fit(X)
# Apply the scaler to X
#scaled_X = X_scaler.transform(X)
# Define the labels vector
y = np.hstack((np.ones(len(car_features)), np.zeros(len(notcar_features))))
# Split up data into randomized training and test sets
rand_state = np.random.randint(0, 100)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=rand_state)
# Shuffle
#X_train, y_train = shuffle(X_train, y_train)
#X_test, y_test = shuffle(X_test, y_test)
print('Using:',orient,'orientations',pix_per_cell,
'pixels per cell and', cell_per_block,'cells per block')
print('Feature vector length:', len(X_train[0]))
Train a classifier¶
In [10]:
# Use a linear SVC
svc = LinearSVC()
# Check the training time for the SVC
t = time.time()
svc.fit(X_train, y_train)
t2 = time.time()
print(round(t2-t, 2), 'Seconds to train SVC...')
# Check the score of the SVC
print('Test Accuracy of SVC = ', round(svc.score(X_test, y_test), 4))
# Check the prediction time for a single sample
n_predict = 20
print('SVC predicts: ', svc.predict(X_test[0:n_predict]))
print('For ', n_predict, 'labels: ', y_test[0:n_predict])
Utility functions todrawing windows on test image¶
In [11]:
# Define a function to draw bounding boxes
def draw_boxes(img, bboxes, color=(0, 0, 255), thick=6):
# Make a copy of the image
imcopy = np.copy(img)
# Iterate through the bounding boxes
for bbox in bboxes:
# Draw a rectangle given bbox coordinates
cv2.rectangle(imcopy, bbox[0], bbox[1], color, thick)
# Return the image copy with boxes drawn
return imcopy
Function to find cars in image¶
Modified function from lesson
In [12]:
# Define a single function that can extract features using hog sub-sampling and make predictions
def find_cars(img, ystart, ystop, scale, cspace, svc, orient, pix_per_cell, cell_per_block, hog_channel, print_all=False):
rectangles = []
img = img.astype(np.float32)/255
xstart=100
xstop=1280
img_tosearch = img[ystart:ystop,xstart:xstop,:]
# apply color conversion if other than 'RGB'
if cspace != 'RGB':
if cspace == 'HSV':
ctrans_tosearch = cv2.cvtColor(img_tosearch, cv2.COLOR_RGB2HSV)
elif cspace == 'LUV':
ctrans_tosearch = cv2.cvtColor(img_tosearch, cv2.COLOR_RGB2LUV)
elif cspace == 'HLS':
ctrans_tosearch = cv2.cvtColor(img_tosearch, cv2.COLOR_RGB2HLS)
elif cspace == 'YUV':
ctrans_tosearch = cv2.cvtColor(img_tosearch, cv2.COLOR_RGB2YUV)
elif cspace == 'YCrCb':
ctrans_tosearch = cv2.cvtColor(img_tosearch, cv2.COLOR_RGB2YCrCb)
else: ctrans_tosearch = np.copy(image)
if scale != 1:
imshape = ctrans_tosearch.shape
ctrans_tosearch = cv2.resize(ctrans_tosearch, (np.int(imshape[1]/scale), np.int(imshape[0]/scale)))
ch1 = ctrans_tosearch[:,:,0]
ch2 = ctrans_tosearch[:,:,1]
ch3 = ctrans_tosearch[:,:,2]
# Define blocks and steps as above
nxblocks = (ch1.shape[1] // pix_per_cell) + 1
nyblocks = (ch1.shape[0] // pix_per_cell) + 1
nfeat_per_block = orient*cell_per_block**2
# 64 was the orginal sampling rate, with 8 cells and 8 pix per cell
window = 64
nblocks_per_window = (window // pix_per_cell) - 1
cells_per_step = 2 # Instead of overlap, define how many cells to step
nxsteps = (nxblocks - nblocks_per_window) // cells_per_step
nysteps = (nyblocks - nblocks_per_window) // cells_per_step
# Compute individual channel HOG features for the entire image
if hog_channel == 'ALL':
hog1 = get_hog_features(ch1, orient, pix_per_cell, cell_per_block, feature_vec=False)
hog2 = get_hog_features(ch2, orient, pix_per_cell, cell_per_block, feature_vec=False)
hog3 = get_hog_features(ch3, orient, pix_per_cell, cell_per_block, feature_vec=False)
else:
hog = get_hog_features(hog_channel, orient, pix_per_cell, cell_per_block, feature_vec=False)
for xb in range(nxsteps):
for yb in range(nysteps):
ypos = yb*cells_per_step
xpos = xb*cells_per_step
xleft = xpos*pix_per_cell
ytop = ypos*pix_per_cell
# Extract HOG for this patch
if hog_channel == 'ALL':
hog_feat1 = hog1[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
hog_feat2 = hog2[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
hog_feat3 = hog3[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
hog_features = np.hstack((hog_feat1, hog_feat2, hog_feat3))
else:
hog_features = hog[ypos:ypos+nblocks_per_window, xpos:xpos+nblocks_per_window].ravel()
test_prediction = svc.predict(hog_features)
if test_prediction == 1 or print_all == True:
xbox_left = xstart+np.int(xleft*scale)
ytop_draw = np.int(ytop*scale)
win_draw = np.int(window*scale)
rectangles.append(((int(xbox_left), int(ytop_draw+ystart)),(int(xbox_left+win_draw),int(ytop_draw+win_draw+ystart))))
return rectangles
In [13]:
test_images = glob.glob('test_images/test*.jpg')
images_pbar = tqdm(test_images, 'Processing', unit='images')
fig, axes = plt.subplots(3, 2, figsize=(16, 14))
axes = axes.ravel()
for i, image_file in enumerate(images_pbar):
image = mpimg.imread(image_file)
draw_image = np.copy(image)
rectangles = find_cars(image, 400, 640, 1.5, colorspace, svc, orient, pix_per_cell, cell_per_block, hog_channel)
window_image = draw_boxes(draw_image, rectangles, color=(0, 0, 255), thick=6)
axes[i].imshow(window_image)
axes[i].set_title(image_file.split('\\')[-1], fontsize=14)
axes[i].axis('off')
Search areas in image frame¶
Defining areas for call find_cars function to avoid search in unexpected areas and improve performance. Specifiying different scales also improves reliability of predictions
In [14]:
locations = []
locations.append({ 'ystart': 400, 'ystop': 465, 'scale': 1.0 })
locations.append({ 'ystart': 420, 'ystop': 480, 'scale': 1.0 })
locations.append({ 'ystart': 400, 'ystop': 500, 'scale': 1.5 })
locations.append({ 'ystart': 430, 'ystop': 530, 'scale': 1.5 })
locations.append({ 'ystart': 400, 'ystop': 530, 'scale': 2.0 })
locations.append({ 'ystart': 430, 'ystop': 560, 'scale': 2.0 })
locations.append({ 'ystart': 400, 'ystop': 560, 'scale': 3.0 })
locations.append({ 'ystart': 460, 'ystop': 660, 'scale': 3.0 })
locations.append({ 'ystart': 400, 'ystop': 600, 'scale': 3.5 })
locations.append({ 'ystart': 465, 'ystop': 660, 'scale': 3.5 })
locations = np.array(locations)
image = mpimg.imread('test_images/test1.jpg')
window_img = np.copy(image)
rectangle_list = []
for location in locations:
ystart = location['ystart']
ystop = location['ystop']
scale = location['scale']
rectangle_list.append(find_cars(image, ystart, ystop, scale, colorspace, svc, orient, pix_per_cell, cell_per_block, hog_channel, print_all=True))
rectangles = []
for sub_list in rectangle_list:
for rectangle in sub_list:
rectangles.append(rectangle)
window_img = draw_boxes(window_img, rectangles, color=(0, 0, 255), thick=6)
image = mpimg.imsave('output_images/all_search_areas.jpg', window_img)
plt.figure(figsize = (20, 10))
plt.imshow(window_img)
Out[14]:
Heat functions¶
In [15]:
def add_heat(heatmap, bbox_list):
# Iterate through list of bboxes
for box in bbox_list:
# Add += 1 for all pixels inside each bbox
# Assuming each "box" takes the form ((x1, y1), (x2, y2))
heatmap[box[0][1]:box[1][1], box[0][0]:box[1][0]] += 1
# Return updated heatmap
return heatmap
def apply_threshold(heatmap, threshold):
heatmap_threshold = np.copy(heatmap)
# Zero out pixels below the threshold
heatmap_threshold[heatmap_threshold <= threshold] = 0
# Return thresholded map
return heatmap_threshold
def draw_labeled_bboxes(img, labels, color=(0, 0, 255), thick=6):
# Iterate through all detected cars
for car_number in range(1, labels[1]+1):
# Find pixels with each car_number label value
nonzero = (labels[0] == car_number).nonzero()
# Identify x and y values of those pixels
nonzeroy = np.array(nonzero[0])
nonzerox = np.array(nonzero[1])
# Define a bounding box based on min/max x and y
bbox = ((np.min(nonzerox), np.min(nonzeroy)), (np.max(nonzerox), np.max(nonzeroy)))
# Draw the box on the image
cv2.rectangle(img, bbox[0], bbox[1], color, thick)
# Return the image
return img
Testing finding car function on on of test images¶
In [16]:
test_images = glob.glob('test_images/test*.jpg')
images_pbar = tqdm(test_images, 'Processing', unit='images')
fig, axes = plt.subplots(3, 2, figsize=(16, 14))
axes = axes.ravel()
for i, image_file in enumerate(images_pbar):
image = mpimg.imread(image_file)
rectangle_list = []
for location in locations:
ystart = location['ystart']
ystop = location['ystop']
scale = location['scale']
rectangle_list.append(find_cars(image, ystart, ystop, scale, colorspace, svc, orient, pix_per_cell, cell_per_block, hog_channel))
rectangles = []
for sub_list in rectangle_list:
for rectangle in sub_list:
rectangles.append(rectangle)
draw_image = draw_boxes(np.copy(image), rectangles, color=(0, 0, 255), thick=6)
axes[i].imshow(draw_image)
axes[i].set_title(image_file.split('\\')[-1], fontsize=14)
axes[i].axis('off')
plt.savefig('output_images/find_cars_illustration.jpg')
Visualization heat results¶
In [17]:
test_images = glob.glob('test_images/test*.jpg')
images_pbar = tqdm(test_images, 'Processing', unit='images')
fig, axes = plt.subplots(3, 2, figsize=(16, 14))
axes = axes.ravel()
heat_images = { 'name': [], 'image': [] }
for i, image_file in enumerate(images_pbar):
image = mpimg.imread(image_file)
heat_image = np.zeros_like(image[:,:,0]).astype(np.float)
rectangle_list = []
for location in locations:
ystart = location['ystart']
ystop = location['ystop']
scale = location['scale']
rectangle_list.append(find_cars(image, ystart, ystop, scale, colorspace, svc, orient, pix_per_cell, cell_per_block, hog_channel))
rectangles = []
for sub_list in rectangle_list:
for rectangle in sub_list:
rectangles.append(rectangle)
heat_image = add_heat(heat_image, rectangles)
name = image_file.split('\\')[-1]
heat_images['name'].append(name)
heat_images['image'].append(heat_image)
axes[i].imshow(heat_image, cmap='hot')
axes[i].set_title(name, fontsize=14)
axes[i].axis('off')
plt.savefig('output_images/heat_illustration.jpg')
Applying heat threshold and visualize results¶
In [18]:
images_pbar = tqdm(heat_images['image'], 'Processing', unit='images')
fig, axes = plt.subplots(3, 2, figsize=(16, 14))
axes = axes.ravel()
heat_tresh_images = { 'name': [], 'image': [] }
for i, heat_image in enumerate(images_pbar):
heat_tresh_image = apply_threshold(heat_image, 1)
heat_tresh_images['image'].append(heat_tresh_image)
heat_tresh_images['name'].append(heat_images['name'][i])
axes[i].imshow(heat_tresh_image, cmap='hot')
axes[i].set_title(heat_images['name'][i], fontsize=14)
axes[i].axis('off')
plt.savefig('output_images/heat_tr_illustration.jpg')
Make labeling and visualize results¶
In [19]:
images_pbar = tqdm(heat_tresh_images['image'], 'Processing', unit='images')
fig, axes = plt.subplots(3, 2, figsize=(16, 14))
axes = axes.ravel()
label_images = { 'name': [], 'label': [] }
for i, heat_tresh_image in enumerate(images_pbar):
# Visualize the heatmap when displaying
heatmap = np.clip(heat_tresh_image, 0, 255)
# Find final boxes from heatmap using label function
labels = label(heatmap)
label_images['label'].append(labels)
name = heat_tresh_images['name'][i] + " - " + str(labels[1]) + " cars was found"
label_images['name'].append(name)
axes[i].imshow(labels[0], cmap='gray')
axes[i].set_title(name, fontsize=14)
axes[i].axis('off')
plt.savefig('output_images/label_illustration.jpg')
Visualize labeled image¶
In [20]:
labels_pbar = tqdm(label_images['label'], 'Processing', unit='images')
fig, axes = plt.subplots(3, 2, figsize=(16, 14))
axes = axes.ravel()
for i, labels in enumerate(labels_pbar):
image = mpimg.imread(test_images[i])
labeled_image = draw_labeled_bboxes(np.copy(image), labels, color=(0, 0, 255), thick=6)
axes[i].imshow(labeled_image)
axes[i].set_title(label_images['name'][i], fontsize=14)
axes[i].axis('off')
plt.savefig('output_images/final_illustration.jpg')
Function to process one video frame or one image¶
In [21]:
def process_frame(image):
rectangle_list = []
for location in locations:
ystart = location['ystart']
ystop = location['ystop']
scale = location['scale']
rectangle_list.append(find_cars(image, ystart, ystop, scale, colorspace, svc, orient, pix_per_cell, cell_per_block, hog_channel))
rectangles = []
for sub_list in rectangle_list:
for rectangle in sub_list:
rectangles.append(rectangle)
heatmap_img = np.zeros_like(image[:,:,0]).astype(np.float)
heatmap_img = add_heat(heatmap_img, rectangles)
# Apply threshold to help remove false positives
heat_tresh_img = apply_threshold(heatmap_img, 1)
# Find final boxes from heatmap using label function
labels = label(heat_tresh_img)
labeled_img = draw_labeled_bboxes(np.copy(image), labels, color=(0, 0, 255), thick=6)
return labeled_img
In [22]:
test_images = glob.glob('test_images/test*.jpg')
images_pbar = tqdm(test_images, 'Processing', unit='images')
fig, axes = plt.subplots(3, 2, figsize=(16, 14))
axes = axes.ravel()
for i, image_file in enumerate(images_pbar):
image = mpimg.imread(image_file)
draw_image = process_frame(image)
axes[i].imshow(draw_image)
axes[i].axis('off')
Function to process one video frame¶
Additionally storage for historical rectangles (like lane positions in Advanced Lane Lines project) Add also lanes detection from project, mentioned above.
In [23]:
max_hist_rects = 20
def add_rects_to_history(hist_rects, new_rects):
hist_rects.append(new_rects)
if len(hist_rects) > max_hist_rects:
hist_rects = hist_rects[len(hist_rects) - max_hist_rects:]
return hist_rects
def process_frame_video(img):
global hist_rects
rectangle_list = []
for location in locations:
ystart = location['ystart']
ystop = location['ystop']
scale = location['scale']
rectangle_list.append(find_cars(img, ystart, ystop, scale, colorspace, svc, orient, pix_per_cell, cell_per_block, hog_channel))
rectangles = []
for sub_list in rectangle_list:
for rectangle in sub_list:
rectangles.append(rectangle)
rectangles = np.array(rectangles)
if len(rectangles) > 0:
hist_rects = add_rects_to_history(hist_rects, rectangles)
heatmap_img = np.zeros_like(img[:,:,0])
for rect_set in hist_rects:
heatmap_img = add_heat(heatmap_img, rect_set)
#Specify dynamically threshold
heatmap_img = apply_threshold(heatmap_img, 1 + len(hist_rects) // 2)
labels = label(heatmap_img)
lanes_img = lanes.process_frame(np.copy(img))
draw_img = draw_labeled_bboxes(lanes_img, labels, color=(0, 0, 255), thick=6)
return draw_img
Run a test video processing pipeline¶
Running extended function with historical array of rectangles
In [24]:
hist_rects = []
lanes.left_line = lanes.Line()
lanes.right_line = lanes.Line()
lanes.startedFromScratch = True
lanes.maxErrors = False
lanes.fail_iterations = 0
lanes.iteration = 0
test_video_out_file = 'output_videos/test_video.mp4'
clip_test_video = VideoFileClip('test_video.mp4')
clip_test_video_out = clip_test_video.fl_image(process_frame_video)
%time clip_test_video_out.write_videofile(test_video_out_file, audio=False)
In [25]:
HTML("""
<video width="960" height="540" controls>
<source src="{0}">
</video>
""".format(test_video_out_file))
Out[25]:
Run a project video processing pipeline¶
In [26]:
hist_rects = []
lanes.left_line = lanes.Line()
lanes.right_line = lanes.Line()
lanes.startedFromScratch = True
lanes.maxErrors = False
lanes.fail_iterations = 0
lanes.iteration = 0
proj_video_out_file = 'output_videos/project_video.mp4'
clip_proj_video = VideoFileClip('project_video.mp4')
clip_proj_video_out = clip_proj_video.fl_image(process_frame_video)
%time clip_proj_video_out.write_videofile(proj_video_out_file, audio=False)
In [27]:
HTML("""
<video width="960" height="540" controls>
<source src="{0}">
</video>
""".format(proj_video_out_file))
Out[27]:
In [ ]: